Skip to content
How-toAdmin

Use Dailybot in CI/CD

Use Dailybot in CI/CD

A CI/CD job can use Dailybot to notify a chat destination, record an agent progress update, submit supported data, or trigger a saved workflow. Keep the integration small: choose one supported action, authenticate through the pipeline’s secret manager, and decide whether a Dailybot failure should block the job.

Choose API, CLI, or the agent skill

  • Use the public API when the pipeline already sends HTTP requests and you want direct control over a supported endpoint.
  • Use the CLI when the job already installs command-line tools or the same command should run locally and in CI.
  • Use the agent skill pack when an AI coding agent is running the task and should select a supported Dailybot workflow from the installed skills.

The CI/CD recipes contain maintained examples. Use the API reference or CLI reference to confirm the current contract before copying a command into a pipeline.

Store authentication as a secret

Create or select the credential described in API authentication. Add it to the CI provider’s encrypted secret store and expose it only to the job that needs Dailybot.

Do not place a credential in a workflow file, shell snippet, build log, artifact, or pull request. Restrict who can edit the job and rotate the credential if it is exposed.

Pick a clear pipeline event

Connect the Dailybot action to a meaningful state, such as a completed deployment, a failed build, a scheduled report, or a release approval. Include enough context for the recipient to act, such as the repository, environment, job result, and a link back to the run. Do not include secrets or full payloads from the pipeline.

For chat notifications and interactive approvals, start with the Messaging API. For progress reports from automated agents, use the agent capabilities documented in the developer portal.

Decide whether the notification is blocking

A coordination notification usually should not turn a successful deployment into a failed deployment. An approval or policy check may need to block. Make that choice explicit in the job rather than relying on a default retry or error behavior.

After setup, test both the successful action and the failure path. Confirm that the message or report reaches the intended destination and that a Dailybot error affects the pipeline only as designed.